1

Interconnector Economics: Electricity Ralph Turvey

Some definitions and descriptions An interconnector, in the case of electricity, is a cable or overhead line connecting

two separate control areas each with its separate system operator. An interconnector has a

Rated Capacity, which generally varies with ambient temperature and other conditions.

Limitations on one or both of the two transmission networks with one or more

interconnectors can cause Total Transfer Capacity, i.e. the maximum continuous

programmed power exchange between two areas consistent with the safe operation of

both interconnected systems, to fall short of the rated capacity of the interconnectors. Its

estimation starts with a base case load flow analysis performed for a given scenario

relating to a generation schedule and consumption pattern (and hence programmed

exchanges) for all the interconnected networks. The Total Transfer Capacity in each

direction between two areas is then computed by incrementing generation in one of them

and decrementing generation in the other, ceteris paribus, until the security limits in one

of the two systems or the interconnectors themselves are reached. Net Transfer Capacity

can then be determined as lower than total transfer capacity by the amount of a

Transmission Reliability Margin. This takes into account the many uncertainties afflicting

the estimation of total transfer capacity. Available Transfer Capacity is Net Transfer

Capacity less Already allocated capacity, the total amount of transmission rights that have

been allocated. These definitions, which constitute important indicators for market

participants planning cross-border transactions and for system operators to manage these

exchanges, relate to programme values, not to physical flows. (ETSO 200100 a)

In the absence of any limitations in the transmission networks that they connect,

DC interconnectors can operate at their rated power. Power flows over them are directly

controlled by their interconnector operators, so being essentially independent of the nature

of the transmission systems they link and their characteristics and patterns of generation

and demand within them. They will never be subject to unforeseen increases of power

flow due to the outage of system components. Export through a DC interconnector can be

2

treated as a pre-scheduled load in the exporting system and the import can be treated in

the same way as a must-run generator in the importing system, each being the aggregate

of exchanges notified in advance. In consequence, with a single DC interconnector, in

contrast to AC systems1, the physical transmission path is identical with the contract path

and it is not necessary to reserve a capacity margin; its net transfer capacity equals its total

transfer capacity.2

AC interconnectors can only be used between the two systems with synchronous

frequencies. Adjusted for losses, the physical flow over a single such interconnector

equates with the equal and opposite load/generation balances in the two systems. This

flow, like the flows over the transmission networks which it connects is determined in

accordance with physical laws by the whole pattern of loads and generation in the two

systems. If ex ante calculations of these flows show that they would infringe some

security constraint, the correction required may alter the interconnector flow. It may, in

any case, comprise inadvertent components unless both system operators successfully

balance their systems.

Costs and benefits

Their nature There is little to be said in general terms about estimating the cost of a proposed

new interconnector. The uncertainties of cost estimation are the same as with many other

kinds of large engineering project, but, except perhaps for uncertainties about obtaining

the necessary planning permissions for possible routes, they are of the second order as

compared with the uncertainties afflicting estimation of the benefits.

It will be necessary to identify any limitations in the connected networks that may

require the interconnector’s total transfer capacity to be sometimes constrained to below

its rated capacity. This involves finding the critical credible contingency, normally the

fault outage of a line, transformer or generation unit outage, the consequence of which

would require the flow into or from the interconnector to be reduced such that no security

constraints are breached in the interconnected systems and on the interconnector itself.

1 Except for possible Flexible AC Transmission Systems, known as "FACTS" 2 Brazil's Itaipu HVDC Transmission Project is the world's most impressive HVDC interconnector.

It has a total rated power of 6300 MW and a world record voltage of ±600 kV DC. It provides an asynchronous connection between the 50 Hz 12,600 MW Itaipu hydropower plant and the 60 Hz network in São Paulo.

3

This is done by load flow studies, usually assuming peak load conditions. In the case of

the North-South Irish interconnector, for example, at times of high load in the Dublin

area, interconnector exports have to be constrained to ensure security in the event of a

contingency occurring on that part of the Irish Republic’s transmission system.

Reinforcement of the transmission capacity in the area from Dublin to the border would

allow greater use of the interconnector at such times.

The cost of any augmentations to existing networks feeding into and out of an

interconnector which are necessary to relieve any such constraint can be investigated.

Augmentation can take various forms, not only upgrading lines and transformers but also

the installation of reactive plant or devices which automatically reduce the power flow on

the interconnector on detection of a specific event. The resulting extra benefits from the

interconnector must be weighed against the cost of the augmentation.

The benefits from a new interconnector will comprise cost savings from one or

more of the following:

• Deferral of investment in generation;

• Reduction in unserved energy (which can be valued at the value of lost

load);

• Reduction in fuel and other variable operating costs (net of transmission

and distribution losses) by substitution of cheaper generation for more

expensive generation; and

• Reduction in the cost of frequency control, spinning reserve and other

ancillary services costs.

In addition, further net benefits may be secured if an interconnector reduces the

market power of generators in the importing area. One of the objectives of the Tasmania-

Victoria interconnector, Basslink, for example was stated to be to introduce competition

into electricity supply in Tasmania, enabling energy prices to be set by competition rather

than regulation.

There will be gains and losses in producer and consumer surpluses if the net

transfer capacity of an interconnector is sufficiently large for flows through it to have an

influence on prices in either of the two systems. These will be additional to the benefits

from reductions in fuel and other operating costs. A simple long-run static one-period

theoretical model assuming fully competitive behaviour serves to illustrate the

4

relationship between these savings in fuel and other variable cost resulting from an

interconnector and such changes in producer and consumer surplus.

Since total variable cost saving in both power systems from the introduction of an

interconnector is the change in the areas under the rising marginal cost curves, producer

surplus rises by A+B+C in the low-cost exporting system on the left and falls by D in the

high-cost importing system on the right. Consumer surplus falls by A+B in the former and

rises by D+E+F in the latter. Hence there is a net gain of C in the former and of E+F in

the latter, together with some redistribution from consumers to producers and from

producers to consumers respectively. This is in addition to the major gain accruing to the

owners of the rights to use the interconnector, namely the import volume times the import

price minus the export volume times the export price. However, in the extreme case of an

interconnector large enough to wipe out any price difference in excess of transmission

losses, the changes in consumer and producer surpluses would constitute the whole of the

benefits revealed by such an analysis.

An appraisal of the costs and benefits of a proposed North Sea Interconnector

between Norway, with predominantly hydro generation, and Britain, with predominantly

thermal generation, provides an example of a case where changes in producer and

consumer surplus would add considerably to the direct benefits to the owners of the rights

to use the interconnector.(ECON Analyse, 2003) In wet years such an interconnector

would add to total Norwegian export possibilities, so raising prices, while in dry years it

A CA B

D E F

Ex por t Import

5

would extend the import possibilities, so reducing import costs; in other words there

would be a favourable terms of trade effect in some years. As regards daily variations on

the other hand, such an interconnector would add to Norway’s capability for export using

its hydro capacity, allowing it to earn more on peak power exports.

Merchant interconnectors3 These are built for profit, their owners obtaining the difference between prices at

either end of the interconnector multiplied by the flow, whether or not they dispatch it

independently as can be done with a DC interconnector. In the short-run, if a merchant

interconnector had sufficient capacity to reduce the price difference to zero, its owners

would not gain from using it in this way. Instead, ignoring losses, they would equate

marginal cost at the injection terminal with marginal revenue at the delivery terminal.

Alternatively, they could offer capacity contracts to traders and generators, providing a

more secure income than the uncertain price differentials. As regards the initial

investment decision, the most profitable size is that which would equate the average gap

between this marginal cost and revenue with the marginal cost of capacity, whereas the

social optimum would be larger, equating the price gap with the marginal cost of capacity.

It has been suggested that any benefit from an interconnector which arises from

deferring other investment and from contributing to reliability is an externality, disregard

of which will lead to suboptimal investment in the absence of an agreement with the

system or transmission operator for a contribution to its revenue.

Against these arguments for private enterprise must be weighed the powerful

arguments that the transmission operators may be unenterprising; that the real choice may

lie between a merchant interconnector of lesser capacity than ideally desirable and no

interconnector at all. (Littlechikd,2003)

Appraisal methods Estimating the probable benefits from an interconnector is extremely complicated.

The steps of an appraisal are:

1. Establish possible competitor projects to the interconnection

2. Formulate a range of possible market development scenarios;

3. Undertake an economic analysis, under each of the market development scenarios,

of the interconnection, of any competitor interconnection projects and of one or 3 An interesting discussion is provided in Brunekreeft, 2003.

6

more base cases which exclude a new interconnection, consisting of planned and

contemplated generation and other transmission schemes.

The first of these steps requires constructive thinking by engineers. Often, possible

alternative projects will have been put forward by those proposing the interconnector

under consideration.

Market development scenarios are distinguished from one another by the different

assumptions they make about the future growth of demand in the two systems being

interconnected, about future fuel prices and, sometimes, about future government

environmental policy with respect to emissions and renewable energy.

The third step involves simulations of the operation of the two systems and the

interconnector and its alternatives under each of the development scenarios, employing

Monte Carlo methods or at least postulating a number of alternatives in order to examine

the sensitivity of the results to, for example, the rate of load growth, generator forced

outage rates, water availability (in a hydro system)4, fuel costs and capital costs. It may

also be useful to take past data and calculate what use would have been made of the

interconnector if it had existed. The simulations have to rest upon assumptions about

pricing behaviour on the part of generators. This will not only directly determine the

flows across the interconnector but also the amount and type of new entrants and the

timing of the retirement of old plant, with longer-run consequences for the size of these

flows.

The uncertainties afflicting such computations can obviously be very substantial.

Looking forward over a series of years, from knowledge of the costs of new generating

plants it may be possible to predict trend levels of prices at periods of high and low load

― if prospective expansion of renewables capacity and the nature and effects of future

emissions controls can be allowed for. However, the actual volume of interconnector use

will depend upon actual prices in the two systems, and, as historical data demonstrate,

daily and even hourly fluctuations around the trend can be very considerable, depending

upon a host of transitory factors.

The analysis can therefore become very complicated, requiring the use of

computational models to simulate dispatch and necessitating judgements to be made about

the investment responses to future price developments. As an example of such

4 In Norway where power supplies are 99% water based, changes in precipitation can reduce the annual production of electrical energy by 20-30%

7

judgements, one study (Intelligent Energy Systems, 2000) postulated an annuitised entry

cost for open and closed cycle gas turbine plants and assumed that new entry would occur

when the time integral over a year of the excess of market price over fuel and operating

costs sufficed to cover that annuitised cost for a typically sized plant. Allowance was

made for the fact that in the event of demand exceeding capacity, so that load shedding

would occur, the market price would be capped at a level determined by the regulator as

the value of lost load. The calculations were performed for more than one discount rate.

The extent and complexity of the analysis will vary very much according to

circumstances. In the case of a possible interconnector between the Republic of Ireland

and Wales it has been estimated that a full-scale feasibility study would cost €8 to €10m.

(DKM, 2003).

The simplest cases arise when the proposed interconnector would link a

permanently high cost centrally planned system with a permanently low cost centrally

planned system. It would allow immediate savings in fuel and other operating costs for

production and/or for the provision of reserve, while in the long run the interconnector

could be expected to substitute directly for capacity in the importing system if one-way

flows were expected to continue to dominate.

When price differences across the interconnector are likely to change sign,

diurnally, seasonally or permanently, altering the direction of flow across the

interconnector, and when there is uncertainty about the pricing behaviour of competing

and sometimes colluding generators, analysis becomes extremely complicated.

Forecasting prices in one system is difficult enough; forecasts of price differences

between them are even more uncertain.

Further complications can be caused by differences in the way the two electricity

markets function and by the possibility of pooling reserves. In the case of the proposed

North Sea Interconnector, for example, to meet the reserve requirements imposed by the

possibility of extremely high peak demands in the Nordpool area, Statnett could avoid

some of the heavy costs of contracting for demand reductions or for reserve capacity

(provided by adding to generating capacity at hydro plants) by contracting for reserve

capacity in Britain. This benefit would be additional to the major benefit resulting from

reduced generation costs. These, in the case of interconnector between a predominantly

hydro and a predominantly thermal system have two roots. One is that thermal generation

can vary between wet and dry years inversely with the availability of water; the other

arises when one system is capacity-constrained and the other is energy-constrained, the

8

former exporting off-peak, the latter on-peak. (Turvey and Anderson, 1977) Tasmania and

Victoria provide another example. Hydrological risk represents the major electricity

supply uncertainty in Tasmania, where storage constraints constrain the capability to

provide increased energy generation, while installed capacity of 2,263 MW exceeds the

system’s average long-term energy supply capability of about 1,105 MW, taking into

account average water inflows. Most of generation in Victoria is from brown coal stations

(Basslink Joint Advisory Panel, 2002)

The benefits to be gained from price differences across an interconnector may be

divided into two parts in the case of the interconnection of two pools (whether net or

gross): those arising from (i) the pre-gate-closure inter-system transactions of generators

and suppliers, and (ii) cost savings in balancing transactions undertaken by closely

collaborating system operators. The latter can scarcely be estimated in advance for a

period of years, however elaborate the simulations carried out. Opportunities for them will

arise for balancing transactions across the interconnector in either direction when it is not

used to its net transfer capacity, and only in the opposite direction to the flow when it is.

In the particular case of the proposed North Sea Interconnector, for example, the lower

cost of upward regulation in the Norwegian hydro system compared with the higher level

and the volatility of System Buy prices in Britain implied that there would be profits to be

made from Norwegian supplying balancing energy in many hours each year.

Interconnector utilisation

Approximating the ideal An optimal use of a link would be achieved if a single system operator scheduled

operation and dispatched (in a gross pool) or redispatched (in a net pool5) both of the two

5 A pool can be called a "market" if prices are determined by it that equate supply and demand; in

the absence of a pool a network of bilateral transactions and/or a power exchange constitute the market. In a gross pool, all generators sell and all suppliers buy all their energy through the pool; in a net pool they notify bilaterally contracted generation and sales and provide bids for generation increments and demand decrements and offers for generation decrements and demand increments. . The operator of a day-ahead gross pool, but not of an hour-ahead one, can determine what plant is to run (unit commitment) as well as scheduling the outputs of those plants that will run. These outputs may subsequently be modified ("redispatched") in an hour-ahead market. In the absence of a day-ahead market the operator running a gross pool will dispatch the system or, if running a net pool with generator and supplier notification of their plans, will redispatch the system. It is possible for a market for operating reserves to operate in parallel with the energy market, these markets being cleared

9

systems linked by it. An example is provided by the interconnector between North Island

and South Island in New Zealand. In such cases the resulting security-constrained

optimisation of generation would determine optimal flows across the link, taking account

of its rated capacity in just the same way as the rated capacity of all the other transmission

lines. The optimisation would entail reviewing the acceptability of the possible effects of

a number of plausible single or dual contingencies including both outages of the

interconnector and outages of the transmission connected to it. Thus the total southward

transfer capacity of new Zealand’s interconnector is little more than half its northbound

capacity because, with southbound transmission, sufficient instantaneous reserve needs to

be available to avoid a risk of cascade failure of the South Island grid in the event of an

interconnector failure and such reserve is both expensive to provide and limited in

quantity.

We now consider how to approximate these results as well as possible,

concentrating on the case of a single interconnector linking two separate systems with

separate control areas, each with its system operator. We consider solutions which fall

short of establishing some kind of combined day- or hour-ahead market run jointly by the

different system operators. Studies have been made of how this might be achieved in the

North East United States. (LECG, LLC and KEMA Consulting, Inc, 2001)

The problems of designing good arrangements are fewer with a DC interconnector

than with an AC interconnector. Paired generation and demand offers into the markets at

each end of the interconnector made by participants within the net transfer capacity of the

interconnector, if accepted by or on behalf of both of the two system operators, would

produce a firm interconnector transfer schedule before gate-closure. They would then

achieve balancing separately. The net transfer capacity of the interconnector may, as

already noted, be determined as falling short of its rated capacity both on account of

transmission constraints in one or other of the systems it connects6 and by the amount of a

simultaneously to minimise the total cost of meeting demand in these markets. In all cases real-time "balancing" or "regulation" instructions will have to be given by the system operator.

6 Another example is that although the Moyle interconnector between Northern Ireland and Scotland has a rated capacity of 500 MW (2 lines of 250 MW), the transfer capacity of this link is 450 MW due to constraints in Scotland. Again, interconnection capability southward from Scotland was increased to 2200MW by completion of the Cleveland to North Yorkshire Line. Such limitations below rated capacity of an interconnector depend on the design of its associated controls, the state of the power system at each end, including the system load at a particular time, and the direction of power flow. Capability may change with time in accordance with changes in the operating state of the transmission network at each end.

10

transmission reliability margin. If the flows on the transmission system of either of them

regularly infringed the net transfer capacity limit, it should be reduced; if it did so only

under occasional circumstances that system operator could allow for it in the same way as

for any other must-run generation or unmodifiable demand which causes problems.

With a single AC interconnector, the problems of designing arrangements to

accommodate planned energy transfers and post gate-closure deviations from them, are

similar, though the flow across the interconnector cannot be directly decided by the

system operators as it is a result of the whole pattern of generation injections and demand

withdrawals within both of their systems. In all cases, of course, the net flow across the

interconnector at any point of time equals the excess of generation over demand and

losses in the exporting system, which will exceed the excess of demand and losses over

generation in the importing system by the amount of interconnector losses.

The simplest case To ensure that congestion on their interconnector is dealt with, the two system

operators have to co-operate in some way. The most limited but simplest type of co-

operation occurs when market participants do not undertake cross interconnector

transactions and, as in the case of the Malaysia–Singapore AC interconnector, the two

system operators use the interconnector only to economise on fast reserves. Each

independently seeks to balance its own system, so that the only flows across the

interconnector are inadvertent. (CER, 2003) Interconnector flows arising as the result of

such contingencies as the sudden outage of an important generating unit are thus a result

of pooled reserves. In the case of a DC interconnector, where flows are determined by

system operators and ramp rate constraints may apply, there could be an arrangement

whereby the energy taken is returned later in a scheduled transfer, so that no monetary

transactions are involved.

With closer co-operation, but still without market participants undertaking cross

interconnector transactions, the system operators could engage in cross-interconnector

economy transactions, sharing out resulting cost savings.

An interconnector power flow that is acceptable under steady state operation may cause unacceptable overloads or voltage deviations in the event of a credible contingency, and it may be necessary to operate at lower power flow levels in anticipation of a contingency event. This is consistent with standard operating practice for a transmission grid, where loads under normal operation are not permitted to exceed load levels that could not be accommodated in the event of a credible contingency occurring.

11

Such limited relationships clearly miss out on all the savings and efficiencies

which planned energy transfers resulting from cross-interconnector transactions by market

participants could achieve. ( Commission for Energy Regulation, 2003) For such more

ambitious arrangements, mutually consistent pre gate-closure procedures have to be

designed. The two major problems that have to be solved in designing them are as

follows:

1. Ensuring that both or neither of the system operators accept each cross-

interconnector bid/offer pair or injection/withdrawal notification put

forward by market participants.

2. Ensuring that the acceptances fit within the market participants’ rights to

use the interconnector and, in aggregate, do not exceed net transfer

capacity.

Note that a further problem with unit commitment processes running simultaneously but

separately is the inefficiency, arising from the lack of any mechanism for the system

operators to coordinate day ahead commitment and scheduling so as to minimise the joint

cost of meeting reserve requirements. Ideally this requires simultaneous joint optimisation

of generation and reserve provision. A simpler procedure is to set aside a fraction of

interconnector net transfer capacity to allow for some reserve sharing.

1. Matching trades across the interconnector Consider two systems, each with an independently functioning system operator

which receives bids and offers for each hour or half-hour either to inject and withdraw, or

to provide increments and decrements from notified bilateral trades. Some of these bids

and offers and notified trades will be inter-system, i.e. cross-border, trades. Taking

account of security constraints, each system operator can then act to produce an ex ante

balance of total generation plus any net imports with total demand including any net

exports, allowing for losses and transmission constraints.

Such a process would require the same gate closure time in the two systems and

agreement about the net transfer capacity of the interconnector, but does not necessarily

require each system to have the same kind of trading arrangements. Thus one could have a

gross pool into which all generation and demand is bid and offered, while the other could

have a net pool with bids and offers for increments and decrements from notified bilateral

trades. They could have their own arrangements for dealing with imbalances and their

calculations could take account of dynamic constraints and block bids.

12

A net export or import schedule could be produced by each system operator to

show how the aggregate cross-border flow would vary with market price in its system.

The two system operators could then share information about these schedules, starting an

iterative process to combine ex ante balance within each system with matched aggregate

import and export volumes falling within the capability of the interconnector, taking

account of security constraints.

Even if such co-operation can secure an aggregate balance of exports and imports,

unmatched acceptances by the two system operators of individual market participants’

imports and exports will be a problem. Participants will be exposed to risk. Thus take the

case of a generator in A selling to a supply business or consumer, or anonymously to a

market, in B

• With day ahead unit commitment and scheduling it may get scheduled to

generate but not to deliver, or vice versa, if either its generation bid or its

delivery offer is priced too high. It may then subsequently find that it

subsequently gets an unfavourable market price if it seeks to offset and so

undo the half-trade, for example by selling the generation in its own

market if it has not received a matched delivery acceptance in the other.

Alternatively it can accept the risk of paying a balancing charge in one

system and receiving a balancing price in the other,.

• Otherwise, to ensure acceptance in both system operators’ unit

commitment and scheduling processes it could simply nominate the

amount of energy it wishes to generate and deliver, or bid and offer

extremely low prices. But the prices determined independently in the two

day-ahead markets may turn out to be such as to make the transaction

unprofitable, while undoing it by later counter-transactions in the two

markets may equally entail a loss.

• Without a day ahead market. and so with self commitment, the risks from

bidding in one market and offering in the other are similar. Prices may be

unfavourable or failure to match will lead to imbalance charges which may

entail a loss. This risk will be less if both markets are highly liquid, for

then a participant can be fairly confident that he will pay the going price in

his home market and receive the going price in the other market.

13

Arbitrage transactions for a generator to profit from temporary or short-term

price differences across an interconnector can thus be risky; the same holds for pure

traders. But if a generator in A has a longer term arrangement including a contract for

differences with a supply business or consumer in B, the net cost to the customer will be

the strike price agreed in that contract and this will also be the revenue to the generator. If,

on the other hand, there is a net pool in B, one of them can nominate the amount of the

customer’s energy offtake and the price will be as agreed.

The conclusion is that some of the price risks can be hedged under gross pools

which settle prices that clear the market in both systems and that they can be avoided if

they have similar net pool arrangements where generation and offtake nominations are

accepted and honoured. However, trading can be much more difficult where the two

systems have different types of market, as, for example, in Ireland. (NERA, 2003) But

that is not all. Cross interconnector trades also involve a risk that the interconnector

capacity required by a trade will not be available or that an excessive price will have to be

paid for it.

2 Transmission rights The second problem which arises, even when import and export trades are

matched individually (and, in consequence, aggregatively), is the management of

congestion on the interconnector. It constitutes a problem even if the interconnector is

wholly owned by the owner of one of the two transmission systems and its system

operator treats the far end in the other control area as one of its transmission nodes. The

possibility remains that the nominations for planned matched export and import trades

across it could exceed its net transfer capacity.

Administered allocation Ascertaining demands for interconnector transfers and then scaling them all down

in the same proportion to fit the agreed net transfer capacity is one administrative method.

Another is to allocate capacity on a first come first served basis. Enforcement is much

easier with a DC interconnector. Both methods have been used and both are clearly

inferior in principle to allocations which make some use of price signals. These signals

may relate to charges for using the interconnector or to price differences across the

interconnector. However in the extremely complex case of multiple interconnectors

between multiple control areas, as exist in continental Europe, administrative methods

14

have sometimes had to be used pending development and acceptance of multilateral price-

reflective methods.

Implicit rights 1. Market arrangements

2. et flow equal to the net transfer capacity (allowing for losses).

3. By differences in market clearing prices, “market splitting”. Differences in

such prices between the two systems ensure that, allowing for losses, the

excess of accepted bids over offers in the one market and the excess of

accepted offers over bids in the other both equal the net transfer capacity of

the interconnector. Congestion then results in higher day and hour ahead

prices at the importing end of the interconnector than at its exporting end.

Redispatching through counter purchases is usually done after gate closure as part

of system operators’ balancing operations, just as they deal with transmission constraints

that are internal to their systems. The market arrangements under which they do this may

differ between the two control areas; one or both may not produce market clearing prices.

Market splitting, on the other hand, requires a common market mechanism. If this

includes a day ahead mechanism, appropriate unit commitment decisions can be made,

whereas the time for that is past closer to and after gate-closure. Differences in market

clearing prices can therefore produce a more economical way of constraining

interconnector flows, besides conveying appropriate incentives for the location of

generation and demand.

Redispatching imposes a net cost upon the two system operators, since what the

one pays on accepting bids to increase generation on the import side of the interconnector

will exceed what the other receives on offers accepted to reduce generation on the export

side. Some arrangement for agreeing the net cost of the counterflow transactions and

sharing them is needed. This cost can be recovered either through uplift charges or by

levying a transmission use of system charge on inter-area transactions. In the latter case,

the tariff could be levied to recover actual costs estimated ex post, or calculated ex ante on

the basis of simulation studies, under or over recovery being allowed for and carried

forward.

Split markets, in contrast with redispatching, require that prices clear the market in

both systems. They result in a net revenue from the interconnector, equal to the price

difference times the net flow across it, accruing to the owners of rights to use the

15

interconnector. (The institution of financial transmission rights is a subject discussed

below.) Where the market clearing prices are determined in a gross pool, they can either

be uniform within each control area or vary between nodes, reflecting both transmission

losses and internal transmission constraints. Such Locational Marginal Prices, determined

using optimal power flow methods, are used in parts of the United States and, for

generation, in the Republic of Ireland. Each terminal of the interconnector constitutes a

node for which a price is determined as part of the dispatch process. All transmission use,

not merely interconnector use, is thus priced and financial transmission rights can be used

for all parts of the system where there is congestion.

Financial Transmission rights Where market prices in each area are determined so as to ensure that planned

flows between them do not exceed the net transfer capacity of the interconnector,

congestion will result in price differences which can produce a revenue surplus. The flow

across the interconnector will be worth more at one end that at the other. Rights to this

revenue can be sold to market participants who wish to sell in one of the markets and buy

or generate in the other. They will thus be able to acquire hedges against the price

differences and the transmission owner can be remunerated by the sale of these rights.

The rights can be of varying terms, perhaps sold in auctions or traded in a secondary

market.

Physical rights Physical rights which market participants have to own in order to trade across an

interconnector constitute an alternative to market coupling along either of the above lines.

They are feasible when there is but a single interconnector. Rights owners scheduling

bilaterals can submit transactions between control areas that have to be accepted in both

markets.

In total they obviously have to add up to no more than the net transfer capacity of

the interconnector. There has to be a scheme for imposing any emergency curtailments

that prove necessary and an accompanying set of physical ramping rights for each control

area. Also, their total has to be adjusted when rated capacity changes seasonally between

winter and summer. As was explained above, the extent to which net transfer capacity

falls short of total transfer capacity may vary through time, depending upon changes in

the availability of different parts of the two transmission networks and upon alterations in

the geographical patterns of demand and generation. The forecasts jointly made by the

16

system operators will gradually become firmer until the time for unit commitment and

scheduling arrives and net transfer capacity is definitively announced ― by 07:00hrs one

day ahead in the case of the French-English interconnector.

However rights are initially awarded, an efficient secondary market can ensure

that they come into the hands of participants most willing to pay for them and, indeed, in

most cases, rights are transferable. Physical rights may initially be awarded to whoever

has provided the equity finance for the construction of an interconnector, they can

subsequently decide whether and how they wish to use or sell them. Otherwise all or

some of the rights can be auctioned, including rights which have not been sold under

long-term contracts and rights not sold through previous auctions. Auctions, which are not

especially complicated where there is but a single link, can be pay as bid or can apply a

market clearing price; a use it or lose it rule is common.

Rights can be for different durations, in either direction and interruptible or not. In

the interests of preserving competition there may be a maximum limit on the capacity any

one participant may acquire. A balance must be struck between the presence of certainty

in the market, achieved by rights of a longer duration, and short term trading which

enhances competition. Yet new supplier entrants will want some security and hence prefer

longer contracts. Uncertainty about the future market system in the connected countries

will make participants want contracts of shorter duration.

An example of auctioned rights is provided by the Moyle interconnector, a

500MW cable link between Scotland and Northern Ireland. The 2003/04 auction was conducted via a multi-stage process. The

available interconnector capacity was divided into equal amounts over two

separate auctions held seven days apart. Any unsold capacity in the first

auction became available in the second iteration (by the same product type).

The process left open the possibility of a third, or residual auction if the

auction did not clear in round two. A pay-as-bid rule applied to the auctions.

The multi-stage approach was decided upon to allow bidders the

opportunity to correct/reassess their bids following the disclosure of the

average price bid in the first iteration. A time period of one week was given

between the first and second iterations to allow bidders to revisit their bidding

strategy and reassess the market. 1.8 Products of 3, 2 and 1 year duration

were auctioned, along with an interruptible product. Three-year capacity,

auctioned the previous year, was still in place thereby reducing the total

amount of capacity available for auction. (OFREG, 2003)

17

Rights to use the French-English interconnector are auctioned annually,

periodically and one day ahead, different categories of use it or lose it capacity rights

being put up for auction under a pay as bid system. Each category is characterised by its

direction, England→France or France→England, and by its duration, multiple Contract

Years, one Contract Year, a season or one or more Contract Days. The amount put up for

sale consists of a predetermined amount plus capacity left unsold from longer-term

auctions. Capacity rights in both directions are auctioned on a non-firm basis, with users

directly responsible for any trading consequences as a result of changes to interconnector

capacity. Thus they are sold on the basis of a target level of availability (100% for day-

ahead sales) with rebates paid by the system operators or additional capacity payments

paid by users depending on whether the actual availability exceeds or falls short of the

target. The rights apply at mid-Channel, with 2.34% allowed for losses, which must be

taken account of by users in their nominations and their contracts in both countries.

Owners of rights for one year or more can transfer them in either of two ways.

Firstly, they can reassign all or part of them to another user for one or more calendar

months. (They must still pay the Operators for the reassigned capacity but, of course, can

recover their payments from their counter party. Secondly, if they do not intend to use

their capacity they can ask National Grid and RTE to reallocate that capacity to another

user in annual, periodic and (two-days-ahead) daily auctions after previously unallocated

capacity has been auctioned but before capacity lost under the use-it-or-lose-it rule is

auctioned. Again, they are still required to pay for their capacity but, if the reallocation

capacity is sold will receive the auction proceeds, less a fixed fee (initially nil) and

adjusted for curtailment reconciliation payments. Furthermore, to the extent that users

indicate that capacity will be unused, and subject to outages, the system operators will

make the capacity available in the daily auction.

How well can a physical rights arrangement secure economic efficiency in

scheduling and dispatch? Users of the French-English interconnector with rights longer

than a day must give the system operators two days’ notice of their intended level of use

of the Interconnector (taking into account reassignment, reallocation, contract volumes

and bids/offers). This allows unused rights to be made available in the daily auction, but

does introduce a certain rigidity. However this is recognised; National Grid and RTE are

seeking to develop a collaborative procedure to facilitate bids and offers being made in

the respective balancing mechanisms of England & Wales and France. For a bid or offer

18

from a user to be accepted in one market, the system operator concerned will need the

consent of the other. (National Grid Company, 2003, paragraph 2.4.)

Superpositioning Where there are interconnector use rights in both directions, it becomes possible

for the larger of the two opposing sets of trades to exceed net transfer capacity by the

amount of the smaller set of trades. Such netting off against each other of nominations in

opposite directions is known as superpositioning. Without it, trades in the minority

direction at times of congestion would result in an uneconomic underutilisation of the

interconnector. If auctions for each of the two directions of flow are conducted

simultaneously, the netting off can be a result of the auction. But, whether or not rights

are auctioned or otherwise allocated, firm rights in excess of capacity in the majority

direction can only be granted if trades in the opposite direction are firm, that is to say that

they are obligations. If they are only options, but are subject to use it or lose it terms, a

day-before auction, or resale by the interconnector administrator may allow the rights to

be transferred to another market participant who will use them. With easy transferability

of the rights, transfer by means of a bilateral contract may also be possible. Otherwise any

excessive rights will have to be interruptible, their exercise being contingent upon the

rights in the minority direction being actually utilised.

If he exercise of physical rights requires participants to submit a fixed

transmission schedule a day or more in advance, transferability will be enhanced.

However if participants want to transfer an energy flow which varies through the day,

thus making available excessive rights which vary oppositely through the day, both

netting off and transfer of excesses may prove too complicated, so that rights can be

exercised only in respect of blocks of energy. In the case of the French-English

interconnector, where the minimum duration of rights is one day but Users may wish to

vary their usage through the course of the day, superposition capacity is currently

not made available.

Transmission charges Access to a facility whose capacity is limited can, in principle, be regulated by

imposing charges for its use which limit the demand. Such charges for use of an

interconnector, would have to be fixed in advance in order to influence market

participants’ behaviour appropriately. This is scarcely practicable when price differences

across the interconnector are not fairly stable. Charges fixed well in advance for extended

19

periods at levels which would prevent demand from exceeding net transfer capacity at

time of potential maximum flow, would obviously lead to gross underutilisation of the

interconnector at other times.

Circumstances could exist in which the total transfer capacity and hence the net

transfer capacity of an interconnector could be varied at a short run marginal cost. These

would arise if, firstly, there was congestion on the part of a transmission system that feeds

or takes from an interconnector and, secondly, this congestion could be ameliorated by

redispatching. The system operator could then accommodate larger interconnector

transfers and should levy a charge reflecting the marginal cost of the redispatching upon

all trades in the dominant direction. 7

Post Gate-closure Forced outages, demand forecast errors, deviations of generator output from

schedules, and within the hour economic commitment/decommitment of fast start

resources or dispatchable loads collectively can cause deviations from schedule. Some of

these real-time developments may, if not offset, raise flows across the interconnector

above its net transfer capacity or lower that capacity. In either case a counterflow is

needed: a reduction of generation in the exporting system and an increase in the importing

system. This requires joint or co-ordinated real-time action by the two system operators.

Even when arrangements to such action have been in place by the two system

operators, all the possible savings that would be achieved if a single operator dispatched

or redispatched both systems linked may not be achieved. It could well happen that there

would be times when the interconnector was not congested but the real-time market

clearing prices differed between each end of it. Post gate-closure arbitrage to take

advantage of the possible production cost savings thus available being impossible or

limited for market participants, taking full advantage of them would require even closer

collaboration between the two system operators to conduct such arbitrage operations. It

would involve adjustment of interchange schedules by the ISOs in real-time, to enable

7 The scope for such action will be much higher when there are multiple interconnectors linking

two control areas, for then congestion on them is compatible with some of them operating at less than capacity and redispatching can raise the overall net transfer capacity. The French transmission owner, RTE, does in fact compute such congestion costs and levies them ex post upon the supplementary "commercial" capacity thus made available.

20

lower cost production that is available in one system to serve the load in the other

system.8

The French-English interconnector. The Interconnector "gate closure" is day-ahead, whilst NETA gate closure is 1

hour ahead. At present, the continental markets mostly co-ordinate their international

power exchanges day-ahead and the Interconnector daily auction timescales were

designed for compatibility with this. The capacity not previously sold in the longer-term

auctions and capacity not required is auctioned, with pay as bid. One would expect the

prices for transmission rights paid in these auctions to reflect price differences one day

ahead, and this has tended to be the case. Because capacity is auctioned for a whole day

but users may vary their use during a day, when it is expected that France will export for

part of the day and import for part of it too, capacity in both directions will have a value.

Higher auction prices are marked for France→England rights when English day-ahead

base load prices are higher in relation to day-ahead base load prices on the French power

exchange Powernext, and conversely for England→France. When pressure on the

interconnector’s capacity is foreseen, as for example for January 2nd 2004,

France→England export capacity was fully nominated throughout and the daily auction

price was as high as €42.77 per MW/day. The day-ahead base price of €44.3 in England

and Wales much exceeded the corresponding price in France of €27.3. There were no

import nominations and no import capacity was sold for this day and others like it. For

January 7th 2004, in contrast with January 2nd, day-ahead base prices were higher in

France than in England; French export nominations were low and were exceeded by

import nominations, so that the net flow was England→France. However, the day-ahead

nominations frequently fall short of the interconnector’s capacity .

In the absence of congestion9, the interconnector’s full 2000MW capacity being

available for flows in the dominant direction10 , there have been many such occasions

8 Such collaboration has been termed Virtual Regional Dispatch and has been discussed in ISO New England, Inc. and New York ISO, Inc., May 2003. 9 When internal constraints in the French transmission system limit the Available Transmission

Capacity of the interconnector, Supplementary Commercial Capacity may be provided by the French system operator constraining some plant on and other plant off. The costs that this incurs are recovered from users of the interconnector as a congestion charge. (RTE, 2004)

10 Superposition of rights in the two directions is ruled out. Superpositioning is the netting off of the use of rights in the minority direction against those in the majority direction. The rights in the minority direction must be exercised, i.e. they must be obligations, not options. The two sets of rights must relate to coincident periods.

21

when nominations were made in opposing directions, the net flow being the difference

between the larger and the smaller opposing sums of nominations. The following

diagram shows how, for three days in November 2003, day ahead rights in both directions

were used most of the time, except for the peak evening hours. During these hours, when

prices rose sharply, both on the French Powernext exchange and in England,

nominations almost fully utilised the interconnector to send power from France to

England, even though French day-ahead prices were lower than English. Otherwise, the

relation between the direction of the price difference and the direction of nominations

corresponds to expectations.

French English Interconnector November 5 - 7 2003Half hourly nominations to and from England above and below the X axis

-2000

-1500

-1000

-500

0

500

1000

1500

2000

MW

€ 0.00

€ 10.00

€ 20.00

€ 30.00

€ 40.00

€ 50.00

€ 60.00

€ 70.00Day ahead price

French export nominations French import nominations PowerNext price UKPX price

Statistics for the whole of 2003 showed that for France→England flows valued by

the price difference, negative value flows amounted to as much as 60% of the positive

value flows. Regression of hourly flows against price differences showed only a very

weak relationship in the expected direction from low to higher prices. (Frontier

Economics & Consentec, 2004, Appendix A3) This is very different from what would be

expected according to the theoretical construct introduced above in the discussion of costs

and benefits. On days with no capacity congestion, prices in the two markets would differ

22

only by the cost, including the 2.34% losses, of using the interconnector. But in practice,

imperfect foresight and transaction costs make a difference. UKPX prices (here shown as

an hourly average of two half-hourly prices) and Powernext prices are fixed the day

before, and nominations are made in advance of real time. Users pay transmission use of

system charges.

Multiple interconnectors When there are several interconnectors providing meshed links between several

control areas, the actual power flows in all of them, in accordance with physical laws,

depend upon the configuration of all their transmission systems and upon the power

injected and withdrawn at all nodes. The different control areas, each with an independent

system operator, may have different market systems. In Europe there is a continent-wide

meshed system of 24 countries with synchronised frequency. It consists of some 200,000

km of 400 and 220 kV lines, hundreds of generating stations directly linked to the system,

and thousands of substations. In a planning phase, the European Transmission System

Operators organisation annually estimates indicative transfer capacities for a typical

winter and a typical summer peak period according to recent observed states of the power

system. Subsequently, with revisions as the time-frame shortens, operators determine

allowable transactions between adjacent areas bilaterally, in almost all cases, separately

from the energy market, by undertaking load flow analyses which take as given the

patterns of generation and consumption elsewhere. The ways in which transmission rights

are allocated vary. In determining their totals between adjacent areas, margins have to be

allowed because of uncertainty regarding those patterns; in any case, where the transfer

capacity estimates differ between a pair of systems, the lower figure of the two is adopted.

System operators, who have no direct control over the interconnector flows, then have to

balance their systems, while ensuring that security constraints are observed. The resulting

pattern of physical flows across borders will not, in general, coincide with the cross-

border commercial exchanges notified by the possessors of cross-border transmission

rights. “In fact, it is very common on the UCTE system to have a significant ‘border flow’

in one direction, while the ‘border exchange’ is in the other direction.” (ETSO, 2004)

These arrangements are decidedly sub-optimal. Three possible market solutions

which might achieve co-ordination which would ensure an efficient security-constrained

dispatch have therefore been discussed. The first two were mentioned above in the context

of interconnection between only two control areas. The problems in a multilateral context

23

are similar but very much more complex. Increased generation in one control area

matched by increased demand in another will have impacts upon flows in other control

areas and across other borders as well.

1. Market splitting (EuroPEX, 2003) Net or gross pools or power exchanges

on both sides equate generation plus imports with demand including

exports so as to limit imports and exports to the net interconnection

transfer capacities across borders. Thus whenever net transfer capacity is

fully used, prices will differ between control areas, reflecting this

interconnector congestion.

2. Flowgate auctions (ETSO, 2001 b): Co-ordinated auctioning of rights to

inject power in one control area and withdraw power in another, the

amount of rights awarded being determined as part of the auction process

so as to take into account all cross border security constraints.

3. Co-ordinated Re-dispatching Cost +. Charges for cross-border transit

reflect the cost of the co-ordinated redispatching necessary to limit imports

and exports to the net transfer capacity of interconnections

All three would require ETSO, the European Transmission System Operators

organisation, to play an important role, including computation of transfer capacities, for

which data on line capacities in the European system, together with assumptions or

forecasts of injections and withdrawals at all nodes are necessary.

Net border transfer capacities fall short of the sum of the rated capacities of the

interconnectors linking control areas for three reasons. Firstly, they provide security by

allowing for n-1 contingencies; secondly, when one line reaches its limits the others may

not be fully loaded; thirdly, some of the capacity may be set aside to allow the systems to

share reserve capacity.11

In practice, it is suggested that the computations should be simplified by limiting

the analysis to a simplified system in which the nodes within each control area are treated

as though they constituted a single node and the only transmission lines taken into

account are the interconnectors between control areas. Thus the location of loads and

generation within each control area are implicitly taken as given. However changes within

11 3000 MW of primary frequency response is shared among Union pour la Coordination du

Transport de l’Électricité (UCTE) members proportionately to each control area’s annual energy output; sharing of secondary response is bilaterally agreed between some pairs of control areas

24

the total will alter flows across the interconnectors. There are therefore two reasons why

actual flows may exceed the calculated flows, in some cases breaching the security

constraints:

i) The nodal distribution of loads and generation within control areas differs

from what was assumed.

ii) The aggregate of loads and generation within a control area differs from what

was assumed. Even when the forecast is recent, this can occur because of

unplanned generation or transmission outages or because weather conditions

were imperfectly foreseen.

On the continent of Europe, the international auctioning of rights or price

determination under market splitting would have to be done well in advance of real time,

possibly with market closure on the day ahead. It is therefore necessary to consider how

divergences of either type (i) or (ii) could be dealt with, for some of them would increase

interconnector flow levels above the transfer capacities auctioned or price-rationed. One

or more of the system operators of the interconnected control areas would have to

undertake counter-purchases and sales before gate closure12 (through bilateral contracts

or, if there is one, a short-term power exchange) and, after gate closure, through whatever

balancing mechanism is used in their control areas. How often and how many such

actions would become necessary would depend upon the conservatism with which the

amount of rights to be auctioned or the acceptable import/export flows were determined.

But whereas such conservatism would reduce the frequency and magnitude, and hence the

net cost of counter-purchases, it would result in an uneconomic under use of the

interconnectors at all times except when near-real-time load flow analysis shows that

security is threatened.

Market splitting, which entails implicit auctioning of transmission rights, is

applied by Nordel where several of the interconnectors are DC so that there are hardly any

loop flows on the remainder to create the complications which afflict the meshed systems

of the rest of Europe. In consequence, total transfer capacity on each interconnector

equals the maximum feasible physical flow. But in a meshed system, where loop flows 12 Gate closure is the time after which generators cannot vary their nominated outputs (which

must balance with nominated consumption except in England and Wales) unless instructed by the system operator. The time between gate closure and the start of the period varies between countries, from 1 minute ahead in Sweden, 1 hour ahead in England and Wales, Norway and Netherlands to 3 hours in France. Generators (and sometimes loads) submit their balancing mechanism bids to increase or decrease their output during the period to the system operator on the day ahead or at gate closure.

25

are important, the transmission capacities of different interconnectors cannot be defined

independently of each other. Where such flows are significant, as in the case of the bloc F,

I, CH, D, B and NL, the three co-ordination market solutions would all require the use of

load flow analysis to compute maximum acceptable cross-border power flow limits and

Power Transfer Distribution Factors.

Transfer Capacity assessment requires a multi-country network description and a

“base case” postulated load and generation level at each node in the system, covering

“normal” levels of exports and imports. The effect upon line flows of a small generation

increment13 in one control area and a corresponding load increment in another is then

calculated for a series of outage contingencies so that fulfilment of the system operators’

security requirements can be checked. If they are all satisfied the process is repeated until

a security constraint is reached. The additional power transfer over the base case is then

added to the initial base case transfers to obtain the estimated total transfer capacity.

Uncertainties are treated implicitly by allowing subtraction of a transmission reliability

margin from total transfer capacity to obtain Net Transfer Capacity.

Power Transfer Distribution Factors express the marginal flow on each line in the

aggregate system as a function of marginal increments of generation at each node

assuming a marginal increment of demand at some hub node. The effect on flows of an

increment in generation at node x matched by an increment in demand at node y is the

sum of the Power Transfer Distribution Factors from x to the hub and from the hub to y.14

The estimates of Power Transfer Distribution Factors can be limited to cross-border

flows, at the risk of some error, by treating all the nodes in each control area as one hub

on a set of assumptions about the internal nodal generation and demand pattern within

each control area.

The way in which such calculations would be utilised under each of the three

market solutions listed above are as follows:

1) Market splitting. Contingent incremental cross-border interconnector flows would be

computed daily for a base case set of injections and withdrawals for each of a set of

postulated outages. The maximum flow which could be added to the assumed base

13 The location of generation within each control areas affects the power flows on interconnectors

as well as on internal lines and transformers. The outputs from the calculations therefore depend upon how the increments are assumed to be spread among generators, some system operators assuming that it is according to merit order while others assume equi-proportional changes for all generators

14 This assumes no losses. Allowing for them adds a small complication.

26

case flow across each border without reaching the physical capacity of any of its

interconnectors under any of the postulated contingencies would then be estimated.

The power exchanges within each control area would then equate interior demand

with interior supply plus or minus the acceptable import or export flow so determined.

Thus security constrained dispatching would be implemented by the acceptance of

bids and offers on the power exchanges; price differences between control areas

resulting for times when interconnector capacity would be fully utilised. After power

exchange closure, counter purchasing in a balancing market would be required not

only to deal with constraints internal to each control area, but also to cope with

changes from the conditions envisaged when making the interconnector calculations.

Such changes may arise from generation or transmission outages and from unforeseen

demand developments. A single power exchange or at least very closely linked power

exchanges providing day ahead scheduling and unit commitment followed by co-

ordinated counter purchasing up to and during real-time would be required to produce

fully optimised solutions.

2) Auctions. These would rest upon similar calculations, being held possibly annually

and then repeated monthly or weekly and then daily. The MW of rights auctioned at

each stage would consist of those quantities not already committed in previous

auctions, plus or minus amendments due to revisions in the calculations resulting from

the new information obtained as real time approaches, plus any previously committed

rights put up for sale by their owners. Again, after closure of the daily auctions, co-

ordinated counter purchasing in a balancing mechanism would be required, not only to

deal with constraints within each control area, but also to cope with changes from the

conditions envisaged when making the interconnector calculations.

a) Under the ETSO co-ordinated auction procedure, the auctioned rights would

relate to pairs of control areas and the set of bids maximising the total value of

accepted bids subject to all the cross-border constraints would be selected. The

auction operator would thus take account of the matrix of Power Transfer

Distribution Factors.

b) Under a flowgate procedure (which has been proposed in the USA) the matrix of

Power Transfer Distribution Factors would be published for use by market

participants and the rights auctioned would relate separately to each of the cross-

border constraints. Participants would bid for whatever bundles of rights were

shown to be necessary by the chains of Power Transfer Distribution Factors

27

linking the injection control area with the withdrawal control area of each of their

envisaged transactions.

3) Co-ordinated Re-dispatching Cost +. Prices would be fixed by ETSO for different

daily and seasonal periods for firm and for interruptible MWh rights between each

pair of control areas covered by the scheme. Participants would pay these prices upon

nominating their transactions between each pair of control areas. The prices would be

calculated as the cost differences for given base scenarios between:

a) The costs of optimal redispatching necessary to satisfy cross-border constraints in

the absence of any outage contingencies;

b) A probability-weighted average of the costs of optimal redispatching necessary to

satisfy cross-border constraints if these outage contingencies occur.

In addition to all the data required under alternatives (1) and (2), in order to determine

these prices ETSO would need to know the incremental costs of all redispatchable

generation in all participating control areas in each period.

Although the methodology used by different system operators for calculating net

transfer capacities is fundamentally similar, the assumptions and data used by them vary

considerably.(Institute of Power Systems and Power Technology of Aachen University of

Technology and CONSENTEC, 2001) Estimated net transfer capacities between to

neighbouring control areas made by their two system operators thus frequently differ. The

lower of the two estimates is usually adopted by both of them when scheduling and

dispatching. Unique authoritative calculations for any group of countries would require

detailed agreement about: the concept and quantifications of base case flows, the types of

outage contingencies covered, whether switching operations to preserve security are taken

account of, how thermal line current limits are decided for continuous operation and for

temporary overload, whether voltage limits are considered, what uncertainties are to be

allowed for and how the allowances are made.

Counter purchasing would be necessary under all three market solutions to ensure

that real time dispatch conformed to security requirements. If not achieved by establishing

a single operator for the control areas, combining them into one, optimisation would

demand very close co-operation between the separate system operators. It would

presumably entail an iterative process, with successive exchanges of information on each

system’s load flows, generation and load bids and offers and constraint shadow values

following each system operator’s redispatch. Thus one system operator would provide the

others with constraint costs for its constraints and offers and bids for inter-area dispatch

28

from its generators and loads. Another system operator would then redispatch

its generation to meet its load, taking account of these data. It would then provide the

other system operators with constraint costs for its constraints and bids and offers

available for inter-area dispatch, a process continuing through all the other system

operators and then repeated until the solutions converged.

Various settlement issues would have to be dealt with, “First, as generation could

be dispatched in one control area to manage a constraint in another control area, the

settlements process will need to provide for assigning the cost of this interchange. Second,

and less obviously, each control area could potentially capture in its internal charges for

generation and load a portion of the congestion rents attributable to constraints in the

adjacent control areas. Revenue adequacy of each system operator visa-vis its day-ahead

settlements will require procedures to track the collection of congestion rents and

appropriately reassign these rents.” (Harvey , S.M., et.al. p.19)

Concluding remarks Interconnectors yield undoubted benefits, but whether their likely level would

justify their probable cost is in most cases subject to great uncertainty. Benefits from an

existing interconnector can more easily be maximised when a single system operator is

responsible for both unit commitment and dispatch in the connected areas. When there is

no such single operator, but two, securing optimal utilisation of the interconnector is

difficult, even with a single DC link, as the phenomenon of two-way nominations attests.

Ways of dealing with the vastly more complex problem of achieving an approximation to

optimal use of multiple AC links have been discussed. They all would require hourly or

more frequent iteration between load flow analyses and auctions or power exchanges and

extremely close and continuous co-ordination between system operators.

29

References Basslink Joint Advisory Panel, 2002, Final Panel Report on the Basslink Proposal, Chapter 3.

Brunekreeft, G., 2003, Market-based investment in Electricity Transmission Networks: Controllable Flow, Cambridge Department of Economics

CER (Commission for Energy Regulation), 2003, Market Arrangements for Electricity – Interconnector Trading Principles An MAE Consultation

DKM Economic Consultants, Economic and Social Research Institute and Electrotec Ireland, 2003, Costs and Benefits of East-West Interconnection between the Republic of Ireland and UK Electricity Systems: Report to the Commission for Energy Regulation

ECON Analyse, 2003, Kraftkabel til England? - samfunnsøkonomisk lønnsomhet

ETSO, 2001 a, Definitions of Transfer Capacities in liberalised Electricity Markets

ETSO, 2001 b, Co-ordinated Auctioning: A market-based method for transmission capacity allocation in meshed networks.

ETSO, 2004, Cross-border electricity exchanges on meshed AC power systems

EuroPEX, 2003, Using Implicit Auctions to Manage Cross-Border Congestion: ‘Decentralised Market Coupling’, Paper by presented at the Tenth Meeting of the European Electricity Regulatory Forum

Frontier economics & Consentec, 2004, Analysis of Cross-Border Congestion Management Methods for the EU Internal Electricity Market, Study conducted for the European Commission Directorate General Energy and Transport

Harvey, S.M., Pope, S.L., Buechler, J.P., and Thompson, R.M., 2001 Feasibility Study for a Combined Day-Ahead Market in the Northeast

Institute of Power Systems and Power Technology of Aachen University of Technology and CONSENTEC, 2001, Analysis of Electricity Network Capacities and Identification of Congestion; Final report.

Intelligent Energy Systems, 2000, Application of the ACCC Regulatory Test to SNI. Report to TransGrid,

ISO New England, Inc. and New York ISO, Inc., 2003. Virtual Regional Dispatch Concept, Evaluation, and Proposal a joint Working Paper

LECG, LLC and KEMA Consulting, Inc., 2001, Feasibility Study for a combined Day Ahead market in the northeast, Study prepared for the ISO-New England, New York ISO, and Ontario IMO.

Littlechild, S., 2003, Regulated and Merchant Interconnectors in Australia: SNI and Murraylink Revisited, CMI Project, http://www.econ.cam.ac.uk/electricity/

National Grid Company, 2003, Interconnexion France-Angleterre (France-England Interconnector) IFA User Guide Issue 4.1

NERA, 2003, ROI Interface Study An Interim Report for the IME Group

OFGEM, 2004, Review of electricity and gas arrangements for winter 2004/05, Consultation Document

OFREG, 2003, Interconnector Allocation Arrangements 2004-05, Consultation Paper June 2003.

RTE, 2004, Détermination des capacités commerciales d'échange et des coûts de congestion associés,

Turvey, R. and Anderson, D, 1977, Electricity Economics, chapter 18 "Cost structure in hydro and hydro-thermal systems",